> This only came to light after the study had already been running for a few months. That proves that we can no longer tell for certain unless it's literal GPT-speak the author was too lazy to edit themselves.

Rule 3 of the subreddit quite literally bars people from accusing posts of being AI-generated. I have only visited it a few times in recent times, but I noticed quite a few GPT-speak posts with comments calling it out getting removed and punished.